lambda expressions do not properly generate closures.

[Bradley M. Kuhn]

Lambda expressions currently fail basic closure tests on trunk. The included patch uses a newclosure Parrot opcode to force creation of a closure each time the lambda is evaluated, allowing lambda closure to function properly. This patch causes that part of the functions.py test to pass now.

[Bradley M. Kuhn (bkuhn-ebb)]

BTW, the patch is copyrighted by me and licensed under Artistic-2.0-or-later.

[Bradley M. Kuhn (bkuhn-ebb)]

Based on discussion with pmichaud, I implemented this as a pirop instead. Committed to trunk at r80.

[Allison Randal (allison)]

Bradley M. Kuhn wrote:
> BTW, the patch is copyrighted by me and licensed under Artistic-2.0-or-
> later.

That would mean we can't apply it. Pynie is currently still covered by
the old Perl CLA, and under Artistic 2.0, but it will be switching to
the Python CLA and the Python license in the not-to-distant future. We
have verbal agreement from all existing Pynie contributors to sign a
Python CLA when we reach that point, and would need the same from you.

Allison

[Bradley M. Kuhn (bkuhn-ebb)]

> Bradley M. Kuhn wrote:
>> BTW, the patch is copyrighted by me and licensed under Artistic-2.0-or-
>> later.

Allison Randal wrote at 02:19 (EDT):

> That would mean we can't apply it. Pynie is currently still covered by
> the old Perl CLA, and under Artistic 2.0, but it will be switching to
> the Python CLA and the Python license in the not-to-distant future. We
> have verbal agreement from all existing Pynie contributors to sign a
> Python CLA when we reach that point, and would need the same from you.

Ok, I didn't realize it, and assumed the project was Artistic-2.0. I'm
sure I can probably sign the Python CLA. I'll read it this weekend and
let you know, but I seriously doubt it will be a problem.

Sorry for the confusion. I had asked pmichaud on IRC and he seemed to
think Artistic-2.0 was right.

(BTW, I've been committing to Pynie already, pmichaud gave me access, so
I will strive to make sure the licensing all gets sorted out. I tend to
like to be explicit about licensing of patches in bug trackers, because
it helps find implicit things -- like this issue of the Python CLA.)

--

   -- bkuhn